University of Minnesota
School of Physics & Astronomy

Spotlight

Data Handyman

Dominick Rocco
Dominick Rocco
                                                       

Dominick Rocco is graduate student in the NOvA group working on experimental neutrino physics. Rocco describes his role as that of a general handyman, though he’s more likely to be fixing software problems than putting up shelves. He works on offline software, reconstruction, data and analysis tools. For the last year he has been specifically working on batch processing of a large data making sure the group’s algorithms are running in an organized way so general uses can consume and digest that data.

In the last year Fermilab has developed a lot of modern tools to deal with the vast amount of data NOvA has begun to accumulate.

NOvA detectors started taking data in 2013 with the near detector coming online in August 2014. “We record a gigabyte every minute or so, which is a huge amount when you add it up over a year.” The group is processing all the data offline and running reconstruction software on it. Rocco says that Fermilab has made new tools for putting the data on tape, so that the group can use their data and keep it forever. Tape might seem old-fashioned, but Rocco says that it is the really only affordable solution for large volumes of data. “There’s a giant tape array which is expandable, supported by a many petabyte disc array which is used for a cache. The tapes are all accessed by robots.” This system is all at Fermilab. At Minnesota there are more resources such as a huge batch farm (individual CPUs which can run analysis) and a lot of storage here, so students and postdocs can analyze data on site. Rocco says these resources are easier to use since they are smaller.

In the meantime, Rocco is working heavily on data and Monte Carlo comparisons. The main method for understanding what is going on in the detector is to compare the actual data to simulations to see how they agree. “We are looking at the final bit of disagreement to make it as close to perfect as possible. [At the near detector] We are recording millions of neutrinos as if they were pieces of candy. Having hundreds of neutrino events to look at was unimaginable fifty years ago.” They are using near detector events, because it is close to the beam source and they can get many more samples. At the far detector events are far more rare because beam diffuses as the beam travels 800 km.

In addition to making sure that the results can be recorded and analyzed, Rocco and the rest of the group are also pushing toward first physics results. "There’s a lot of excitement to actually do some physics. The reconstruction algorithms were in their infancy when I joined in 2010, and now they are mature and we are starting to observe neutrinos." Rocco says that the neutrinos they’ve observed already have been beautiful. MINOS (previous generation of Neutrino detectors) was 1 cm strips of scintillator sandwiched between planes steel. "You’d register hits in these, and the muons would go pretty far, but most of the neutrino interactions would be absorbed. With NOVA it’s oil and PVC so the neutrino moves through the detector and you see all these prongs coming out and marvelous detail that you just couldn’t get with previous generation detectors. For years we looked at simulations but they are not real. It is a lot more exciting when you see it in real data." Rocco says to expect the first physics results from NOvA in 2015.